cs-onboard
Pass
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill handles sensitive corporate data (revenue, runway, leadership conflicts) but mitigates risk through explicit 'Privacy Rules' and 'Anonymization' protocols. It strictly forbids sending specific financial or personal data externally and provides guidelines for converting sensitive details into safe descriptors before any API calls.
- [SAFE]: The skill manages a local persistent file (
~/.claude/company-context.md) to store captured context. This behavior is consistent with the skill's primary function and is governed by rules requiring explicit user confirmation for updates. - [SAFE]: The skill's architecture for processing user-provided interview data into a persistent file represents a theoretical surface for indirect prompt injection. However, the risk is minimal given the lack of executable capabilities (such as subprocess calls, exec/eval) and the presence of boundary markers in the generated markdown output structure. Ingestion point: User interview conversation in SKILL.md. Boundary markers: Structured markdown headers and labels. Capability inventory: Local file writing to
~/.claude/directory mentioned in SKILL.md. Sanitization: Privacy filtering and anonymization required before any external transmission.
Audit Metadata